[VRRP] Fwd: RFC5798 - clarification on checksum calculation

[Hermin Anggawijaya <[e-mail]>]

Hello

I am thinking of submitting an errata for RFC 5798 Sec. 5.2.8 to make
it more precise in describing the checksum calculation for each address family.

But reading a few responses here, I gathered that there is no general
agreement as to the original intention of the text,
I believe that for IPv4, the checksum is calculated without
pseudo-header so that it is backward compatible with RFC 3768.

Any other opinions - particularly from original authors ?


Kind Regards

Hermin Anggawijaya



On Tue, May 15, 2012 at 2:41 PM,  <[e-mail]> wrote:
> Hello list,
>
> Has there been any progress with regard to this topic?
> I know that in an implementation of VRRPv3 for IPv4 the checksum
> is calculated without pseudo-header currently, but I believe that
> pseudo-header should be involved in the calculation as same as
> IPv6 case.
>
> Regards,
> --
> Tomohiko Kurahashi <[e-mail]>
>
>
> From: [e-mail]
> Date: Mon Apr 02 2012 20:41:51 JST
>>
>> Forwarded.
>> Any other VRRPv3/IPv4 implementation?
>>
>>
>> Thanks,
>> Tomoyuki
>>
>>
>> Date: Mon, 2 Apr 2012 03:26:15 +0200
>> Subject: Re: [VRRP] RFC5798 - clarification on checksum calculation
>> From: Hermin Anggawijaya <[e-mail]>
>> To: Tomoyuki Sahara <[e-mail]>
>>
>> Sahara-san
>>
>> Thanks for your input.
>>
>> Anyone else with either/other interpretation of the clause ?
>>
>>
>> Thanks
>>
>> On Mon, Apr 2, 2012 at 1:48 AM, Tomoyuki Sahara <[e-mail]> wrote:
>> > Hi,
>> >
>> > On Thu, Mar 29, 2012 at 4:29 PM, Hermin Anggawijaya
>> > <[e-mail]> wrote:
>> >> Would someone be able to help clarifying RFC5798 Sec. 5.2.8 on
>> >> checksum for me please...
>> >>
>> >> It says that
>> >>
>> >>  "The checksum is the 16-bit one's complement of the one's complement
>> >>   sum of the entire VRRP message starting with the version field and a
>> >>   "pseudo-header" as defined in Section 8.1 of [RFC2460].  The next
>> >>   header field in the "pseudo-header" should be set to 112 (decimal)
>> >>   for VRRP.  For computing the checksum, the checksum field is set to
>> >>   zero.  See RFC1071 for more detail [RFC1071]."
>> >>
>> >> My interpretation of the above clause is, for IPv4 VRRP the checksum would be
>> >> defined as:
>> >>
>> >> "The checksum is the 16-bit one's complement of the one's complement
>> >>  sum of the entire VRRP message starting with the version field"
>> >>
>> >> as per RFC 3768, instead of involving "pseudo header" (as defined in
>> >> Section 8.1 of [RFC2460]).
>> >
>> > My understanding is only reference text ("as defined in Section 8.1 of
>> > [RFC2460]") is irrelevant for IPv4.  Our implementation calculates checksum
>> > including pseudo header as for TCP/UDP/DCCP.
>> >
>> >> If my interpretation is correct, would it be useful to change the text to
>> >> reflect specific checksum detail for IPv4 ?
>> >
>> > My interpretation is different from yours but clarification should be
>> > very useful.
>> > It's vital for interoperable implementations of VRRPv3/IPv4.
>> >
>> >
>> > Thanks,
>> > Tomoyuki
> _______________________________________________
> vrrp mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/vrrp